pallet-scheduler: Introduce OriginPrivilegeCmp (#10078)

* pallet-scheduler: Introduce `OriginPrivilegeCmp`

When a scheduled task should be canceled, the origin that tries to
cancel the task is compared to the origin the task should be executed
with. Before this pr this check only allowed that both origins are
equal. However, this is problematic as this means that for example a
council origin it needs to be have the same amount of yes votes to
cancel the scheduled task. While a council origin with more yes votes
should be able to cancel this task. This happened recently on Kusama and
lead to a failed cancelation of a scheduled task. With this pr the two
origins are compared and the cancelling origin needs to have greater or
equal privileges as the origin that scheduled the task. What a greater,
equal or less privilege is, can be configured in the runtime.

For simplicity, a `EqualPrivilegeOnly` implementation is provided that
only checks if two origins are equal. So, this mimics the old behaviour.

* FMT

* fix import

* Small optimizations

Co-authored-by: Shawn Tabrizi <shawntabrizi@gmail.com>

substrate/bin/node/runtime/src/lib.rs

@@ -26,8 +26,8 @@ use codec::{Decode, Encode, MaxEncodedLen};
 use frame_support::{
 	construct_runtime, parameter_types,
 	traits::{
-		Currency, Everything, Imbalance, InstanceFilter, KeyOwnerProofSystem, LockIdentifier,
-		Nothing, OnUnbalanced, U128CurrencyToVote,
+		Currency, EqualPrivilegeOnly, Everything, Imbalance, InstanceFilter, KeyOwnerProofSystem,
+		LockIdentifier, Nothing, OnUnbalanced, U128CurrencyToVote,
 	},
 	weights::{
 		constants::{BlockExecutionWeight, ExtrinsicBaseWeight, RocksDbWeight, WEIGHT_PER_SECOND},
@@ -345,6 +345,7 @@ impl pallet_scheduler::Config for Runtime {
 	type ScheduleOrigin = EnsureRoot<AccountId>;
 	type MaxScheduledPerBlock = MaxScheduledPerBlock;
 	type WeightInfo = pallet_scheduler::weights::SubstrateWeight<Runtime>;
+	type OriginPrivilegeCmp = EqualPrivilegeOnly;
 }
 
 parameter_types! {